An Intelligent Coordinated Control System for Power Transformers Using Deep Q-Network DOI Creative Commons

Ju Guo,

Wei Du, Guozhu Yang

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 108797 - 108809

Published: Jan. 1, 2024

Automatic coordinated control of power transformers is essential to stable operation systems. However, there still lack mature intelligent solutions for this purpose. As a result, paper proposes reinforcement learning-based automatic collaborative approach transformers. Firstly, by establishing the state space and action transformer system, deep learning used optimize strategy. Then, Q-network utilized automatically adjust operating parameters each achieve In algorithm design, we consider multiple factors such as grid load voltage stability. And they are incorporated into reward function facilitate appropriate strategies. The experimental results show that new method not only improves response speed but also effectively enhances stability robustness. addition, conducted in-depth analysis on convergence efficiency verifying feasibility superiority proposed compared with traditional methods.

Language: Английский

Discernment of transformer oil stray gassing anomalies using machine learning classification techniques DOI Creative Commons
M. K. Ngwenyama, Michael Njoroge Gitau

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Jan. 3, 2024

Abstract This work examines the application of machine learning (ML) algorithms to evaluate dissolved gas analysis (DGA) data quickly identify incipient faults in oil-immersed transformers (OITs). Transformers are pivotal equipment transmission and distribution electrical power. The failure a particular unit during service may interrupt massive number consumers disrupt commercial activities that area. Therefore, several monitoring techniques proposed ensure maintains an adequate level functionality addition extended useful lifespan. DGA is technique commonly employed for state OITs. understanding samples conversely unsatisfactory from perspective evaluating relies mainly on proficiency test engineers. In current work, multi-classification model centered ML demonstrated have logical, precise, perfect DGA. used analyze 138 transformer oil (TO) exhibited different stray gassing characteristics various South African substations. combines design four classifiers enhances diagnosis accuracy trust between manufacturer power utility. Furthermore, case reports using model, IEC 60599:2022, Eskom (Specification—Ref: 240-75661431) standards presented. addition, comparison conducted this against conventional approaches validate model. demonstrates highest degree 87.7%, which was produced by Bagged Trees, followed Fine KNN with 86.2%, third rank Quadratic SVM 84.1%.

Language: Английский

Citations

6
Diagnostic and Prognostic Health Management of Electric Vehicle Powertrains: An Empirical Methodology for Induction Motor Analysis DOI
Hicham El Hadraoui, Oussama Laayati, Adila El Maghraoui

et al.

Published: June 14, 2023

The growing interest in electric vehicles has led to an increased focus on the development of efficient and reliable motors. To ensure operation, it is essential incorporate on-board diagnostic prognostic tools that can detect predict potential failures. This paper proposes approach diagnose health condition induction motors used vehicle powertrain applications using machine learning techniques. proposed utilizes vibration signals collected from accelerometers attached motor employs decision forest tree algorithms classify motor. study aims identify most significant features evaluate effectiveness diagnosing predicting models are trained full extracted selected Principal Component Analysis (PCA) Correlation (CA) improve classification performance. experimental results demonstrate combination PCA with Decision Forest (DF) algorithm achieves best performance for simulated fault conditions. suggests techniques be effective applications.

Language: Английский

Citations

7
Effect of Large-scale PV Integration onto Existing Electrical Grid on Harmonic Generation and Mitigation Techniques DOI
Adila El Maghraoui, Younes Ledmaoui, Oussama Laayati

et al.

Published: June 14, 2023

Power quality issues can arise in an electrical grid due to various factors, and one of the most common is harmonic distortion. Harmonics are essentially sinusoidal signals at frequencies that multiples fundamental frequency, they occur when nonlinear loads such as variable speed drives, electronic ballasts, computer power supplies connected grid. The integration large-scale photovoltaic (PV) systems into has led increase distortion, which affect stability reliability In this paper, we use ETAP software analyze impact PV on distortion A model system created ETAP, a analysis performed determine content system. results show generates significant levels To mitigate harmonics generated by system, mitigation techniques analyzed. Passive filters were sized, implemented network, tested using well capacitor banks resonance impact.

Language: Английский

Citations

7
Toward an Intelligent Diagnosis and Prognostic Health Management System for Autonomous Electric Vehicle Powertrains: A Novel Distributed Intelligent Digital Twin-Based Architecture DOI Creative Commons
Hicham El Hadraoui,

Nada Ouahabi,

Nabil El Bazi

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 110729 - 110761

Published: Jan. 1, 2024

Language: Английский

Citations

1
Energy cost forecasting and financial strategy optimization in smart grids via ensemble algorithm DOI Creative Commons

Yang Juanjuan

Frontiers in Energy Research, Journal Year: 2024, Volume and Issue: 12

Published: Aug. 29, 2024

Introduction In the context of energy resource scarcity and environmental pressures, accurately forecasting consumption optimizing financial strategies in smart grids are crucial. The high dimensionality dynamic nature data present significant challenges, hindering accurate prediction strategy optimization. Methods This paper proposes a fusion algorithm for grid enterprise decision-making economic benefit analysis, aiming to enhance accuracy predictive capability. method combines deep reinforcement learning (DRL), long short-term memory (LSTM) networks, Transformer algorithm. LSTM is utilized process analyze time series data, capturing historical patterns prices usage. Subsequently, DRL employed further enabling formulation optimization purchasing usage strategies. Results Experimental results demonstrate that proposed approach outperforms traditional methods improving cost Notably, on EIA Dataset, achieves reduction over 48.5% FLOP, decrease inference by 49.8%, an improvement 38.6% MAPE. Discussion research provides new perspective tool management grids. It offers valuable insights handling other high-dimensional dynamically changing processing decision problems. improvements highlight potential widespread application sector beyond.

Language: Английский

Citations

1
Smart Energy Management System: Predictive Maintenance for Dry Power Transformers Using Transfer Learning DOI
Oussama Laayati,

Oumaima Amaziane,

Mostafa Bouzi

et al.

Published: July 19, 2023

Dry power transformers are a critical component of microgrids, but their diagnostic can be challenging due to the various types defects that occur. This paper proposes several monitoring techniques predict these and improve dry in microgrids. One key features this approach is use thermal image classification detect number short circuits transformer. The images performed using transfer learning method, which allows for utilization pre-trained models adaptation them specific task at hand. feature integrated into larger smart energy management system aims optimize operation maintenance systems. proposed have been tested validated through experiments, results demonstrate effectiveness accurately predicting improving

Language: Английский

Citations

1
Data-driven based Power Quality Disturbance Analysis for Improved Reliability in Smart Grids DOI
Adila El Maghraoui, Younes Ledmaoui,

Ahmed Chebak

et al.

Published: June 4, 2024

Language: Английский

Citations

0
Enhancing Electric Vehicle Diagnostics Through Constant Speed Subrange Detection for Noise-Reduced Analysis DOI
Hicham El Hadraoui, Nasr Guennouni, Adila El Maghraoui

et al.

Published: June 25, 2024

Language: Английский

Citations

0
An Intelligent Coordinated Control System for Power Transformers Using Deep Q-Network DOI Creative Commons

Ju Guo,

Wei Du, Guozhu Yang

et al.

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 108797 - 108809

Published: Jan. 1, 2024

Automatic coordinated control of power transformers is essential to stable operation systems. However, there still lack mature intelligent solutions for this purpose. As a result, paper proposes reinforcement learning-based automatic collaborative approach transformers. Firstly, by establishing the state space and action transformer system, deep learning used optimize strategy. Then, Q-network utilized automatically adjust operating parameters each achieve In algorithm design, we consider multiple factors such as grid load voltage stability. And they are incorporated into reward function facilitate appropriate strategies. The experimental results show that new method not only improves response speed but also effectively enhances stability robustness. addition, conducted in-depth analysis on convergence efficiency verifying feasibility superiority proposed compared with traditional methods.

Language: Английский

Citations

0